How Do Dating App Analytics Work to Match People?

Humans have been flirting and "super liking" since the dawn of humanity. We may not know how our Neanderthal ancestors did it (did they pound their chests like gorillas?), but we do know some more modern attraction techniques involved dropping a fan, showing a tantalizing ankle, prom dates, drive-in movie theatres, and nervously calling their landline praying their parents don't pick up.

In the last few decades, the latest evolution in dating has emerged. Location-based dating applications, expertly crafting an appealing online profile, and swiping right have become the norm. So much has the online dating world altered how we meet that according to a 2013 study, one-third of marriages in the United States are between partners who were introduced online.

How does matching work on dating apps?

Most dating apps keep how they match users largely under wraps. Most online dating platforms use some form of artificial intelligence (AI) to introduce potentially compatible users. Still, the intricacies of it all are hidden, and executives dodge detailed questions on the regular.

However, we do know some things. In the past, dating sites like relied on lengthy questionnaires to determine compatibility.

eHarmony became one of the first to develop and patent an effective matching algorithm, created by a team of psychologists and the company's founder, Dr. Neil Clark Warren. Surveying 5,000 married couples, they used the findings to increase match rate. By today's standards, their algorithm was simple. Using a regression-based approach (essentially comparing multiple variables such as interests, income, sexual preferences, religion, etc, to determine what related factors created the best pairings) to predict long-term relationship satisfaction.

While revolutionary, these algorithms were generally poor indicators of what people actually found desirable and couldn't predict long-term relational success.

The iPhone, quick sign-ups, and chess

The first location-based app, Grindr, was released in 2009, following the release of the iPhone in 2007. As we well know, apps have a rapid sign-up process, and this shift meant developers moved to a collaborative filtering model.

Collaborative filtering algorithms suggest recommendations based on the behaviours of users who appear to have similar tastes.

The Elo System

Used to rank the world's top chess players, Tinder revolutionized the dating game by applying the Elo System to its swiping method in 2012. Taking a pawn from chess' board, Tinder gamified dating by literally mirroring a deck of cards where users could swipe right or left for "yes, I'm interested" or "no, not for me". A double opt-in system, both users must swipe right to confirm the connection.

In chess, The Elo System assigns players a score based on their previous wins and losses and the skill levels of their last opponents. On Tinder, the more right swipes someone had, the more desirable they were. If you swiped right on someone with many likes, you boosted your score.

This raised issues as Tinder is primarily based on first impressions and photos, meaning your desirability was based almost exclusively on how you looked. This disproportionately affected visible minorities and those who didn't fit the made-up "ideal standard" of beauty. Tinder has since ditched the Elo System and instead encourages users to use the app more to improve their algorithm (and so Tinder can gather more user engagement metrics, but they don't mention that part).

The Gale-Shapley algorithm

Also launched in 2012, Hinge uses the Gale-Shapley algorithm. Another form of collaborative filtering, matches are considered stable if no two people would rather be with each other than the person recommended to them.

For example, by matching Lucas with Hannah, either party can be confident that there is no one else in the dating pool they would prefer who would also be interested in them. Obviously, dating apps repeat this process with multiple matches, but each match is apparently exclusive. This algorithm faces the same dilemma as Elo; matches are based on assumed desirability, a poor indicator of compatibility.

How do dating apps determine compatibility?

The short answer is that they don't. Dating app algorithms work broadly on desirability. A 2017 study used a machine learning algorithm to predict romantic desire using relationship science. While the algorithm could pick out factors based on desirability and selectivity, it couldn't predict long-term compatibility.

While algorithm-based apps like okCupid put more emphasis on a user's preferences and, thus, could potentially better predict compatibility than swipe-based apps like Bumble or Hinge, a 2012 (pre-Tinder) study revealed that none of the dating apps could do a better job matching people than the randomness of the universe could. In fact, the best predictor of long-term relational success was "the way they [the two potential matches] respond to unpredictable and uncontrollable events that have not yet happened."

In other words, algorithms aren't the conduit to love; random chance (or fate, if you will) is. Perhaps make like the people on TikTok and start looking lost in a Barnes & Noble or a Home Depot. Some handsome stranger may suggest the latest thriller or help you decide what size hinge (get it?) you need. Boom, love.

However, that same study revealed that dating apps give you access to more people than you might meet daily and facilitate conversations with said people. So, if dating apps are your thing, no harm, no foul.

Do you like, Super Like them?

As a kid, you might have passed a note to your schoolyard crush asking them if they liked you by checking "yes" or "no." Tinder has since changed the question: "Do you super like me?"

But what is the Super Like functionality? Essentially, instead of swiping right on someone you like and hoping they do the same, the Super Like tells your desired swipe that you're into them. You get one free Super Like per day or can pay to have more—either way, Tinder says it triples your chances of getting a match. This is because Super Liking someone puts you outside of the regularly scheduled user behaviour analysis, as the algorithm is now obligated to push your profile closer to the top of the pile of the person you picked.

The question here is: so what? Your Super Like pushes your card to the top of their queue, and they can see you're interested. Cool. You're not guaranteed the match, as they can still swipe left. In fact, the outcome is similar if you stick to swiping right on nine people per day (this is science, people) and then closing the app. Truly successful dating profiles take a dash of effort.

Which dating app has the highest match rate?

According to a study by Currys, Tinder users have the best chance of finding love. However, algorithms are like the human brain: malleable. You can force it to work in your favour regardless of your chosen app with solid profile optimization strategies (aka, not infinitely swiping right).

  1. Funny, creative bios tend to get more matches.
  2. UNLESS IT'S YOUR ACTUAL HOBBY, GET THAT DANG FISH OUT OF YOUR PROFILE PHOTO. I don't have stats to back this up, but I'm 90% certain that people (and I'm mainly talking to the guys here) aren't that impressed. Unless fishing is something you do regularly, that one time you caught a trout the size of my desk lamp isn't going to impress anyone.
  3. Speaking of doing things regularly, include your hobbies (yes, fishing is fine) and interests. If you have an animal, add a snapshot of you with Fluffy and/or Fido.
  4. I've said it before, and I'll say it again, if you post a group photo, tell us which one you are. Include mainly individual snapshots. Group photos are ok as potential matches like to see that you have friends, and group pics can reflect hobbies; just tell them which one you are.
  5. Stay true to who you are. If you say you go to the gym just to get matches, keeping up the ruse when you fall for someone who works out religiously will be hard. Oh, and please stop lying about your height.
  6. Link your Spotify or Instagram; this can help people learn more about you before swiping left or right. Note: If you're worried about online privacy, it's best to skip this option.
  7. Tinder and other apps will release new features from time to time, like specifying the ideal outcome of a match and the chance to add even more personal info about yourself.

Dating the data

It goes without saying that dating apps need to collect your location as it’s unrealistic to match you with someone 300km away (unless you’re that dedicated. You go, Glen Coco). They also collect your name, age, gender, email, likes, sexual preferences, who you’ve liked and who your matches have liked, and even your payment details if you use their premium features.

While many apps allow you to request a data usage report, what measures are they taking internally to protect you and your data?

Staying safe

Like love, most dating app data usage remains a mystery. Researching various security sites, it doesn't seem like dating apps are going above and beyond to protect users' data. This may be because the nature of dating apps is to show yourself off to others via photos and personal descriptions.

Kaspersky conducted a study on dating usage in dating apps and their safety in 2017 and then repeated it four years later to see if specific security measures had improved.

Focusing on the nine most popular dating apps, Tinder, OKCupid, Badoo, Bumble, Mamba, Pure, Feeld, Happn, and Her, the repeated study found some improvements. All nine apps we researched the second time use encryption. All feature a mechanism against certificate-spoofing attacks: if a fake certificate is detected, the apps stop transmitting data. Mamba even gives the user a warning. Although a potential hacker could still access the user's data, this is unlikely given the anti-spoofing mechanism works well against this.

However, things aren't always so rosy. Let's take a closer look.

Two of the nine apps — Mamba and Badoo — send a new user's password in plain text. This is dangerous because many people tend to be lax about password safety. By hacking or intercepting the user's email, an attacker can uncover the password and gain access to their account (unless two-factor [2FA] authentication is enabled in the dating app. Note: ALWAYS activate 2FA if you have the option).

Four apps — Tinder, Bumble, Happn and Her — require location access. All but Happn lets you opt out of exact location tracking and just identify your general region, but only if you have the paid version (hello, credit card details). One app, Mamba, can detect your distance from other users with frightening accuracy: one metre. Giving Mamba your geolocation data is not a good idea.

Overall, Kaspersky determined that dating app security has improved significantly, adding encryption and resisting man-in-the-middle attacks (when the hacker inserts themselves into a supposedly two-way conversation). Many also have bug-bounty programs which reward users for reporting suspicious activity, helping to patch vulnerabilities in their respective systems.

However, apps do little to promote reasonable cybersecurity measures in their users. Most apps encourage oversharing (linking other social accounts, answering question prompts that don't need to be shared with strangers, etc.), ignoring the possible consequences: doxing, stalking, data leakage, and more.

As always, you are the best protector of your data. There are many ways to protect yourself on dating apps, starting with staying informed.

You can use dating apps if you like. Just like debutant balls and the ability to play piano really well in Georgian-era England was a way to get you noticed, so too is swiping right. However, you aren't worse off if dating apps aren't your thing. Putting yourself out there can take on many forms. You can just as easily meet your future life partner at school, in a club, at the gym, at work, while travelling, or anywhere where there are people. Also, if you're conscientious about online digital security, you should stick to more traditional dating methods.

So, fly your Tinder Flag high or just, you know, frequent the same café every week or something.